CSE 691 is a research topics course.
The performance and correctness of complex engineered systems is typically
impacted by many factors, each of which can be set to one of many levels, values, or options.
Once factors and levels are identified, interaction testing seeks to
determine not just the effects of individual levels of factors, but also
interactions among small sets of them.
Because of this, interaction testing has been extensively used in configuration testing for software
and in component-based software design.
It has also been employed in hardware testing, network testing, computational
learning, biological networks, and the design of screening experiments.
In this course we study the theory of interaction test suites:
- lower and upper bounds on their sizes
- explicit combinatorial and algebraic constructions
- randomized, derandomized, and heuristic construction methods
- recursive methods for making large test suites from small ones
- the ability to locate faults
We also explore practical aspects:
- rate of fault detection
- forbidden, optional, and required tests
- test-aware and cost-aware testing
- applications in other areas
Topics can be determined by class interest.
After lectures on August 22, 24, 29, 31, we begin on 7 September to alternate:
On each Monday we have a lecture to introduce further background;
on each Wednesday, we discuss one or two research papers.
The paper(s) will be distributed at least one week in advance.
One student will volunteer (or be volunteered) to lead the discussion, but
all students will read the paper and have formed opinions about them that they can and will express.
My recommendation is that you give each paper a critical reading.
- What does the paper purport to do? Does it do those things?
- Does this appear to be significant? Why or why not? Is it of practical value, or a theoretical advance, or perhaps both?
- Would you recommend it to a friend (who you want to keep as a friend)?
- Is the paper written in a way that makes answering the earlier questions fairly easy? What could have been presented better?
- What does the paper not do, that perhaps it should have done?
- If you were going to try to extend this work, what would you do next?